Exploring the possible risk factors of COVID-19 by combined genetic, functional genomic, clinical, and imaging data
The spread of COVID-19 has become a global pandemic since March 2020, which has caused millions of infections worldwide. We would examine whether COVID-19 morbidity and mortality could be reduced through interventions. It is possible that some risk factors can enhance the predictability power of statistical models. Some subpopulations might be more vulnerable from COVID-19. The project will be undertaken during 36 months.
1. There are many factors associated with COVID-19 infections. We are interested in clinical factors (such as age, the family history of diseases, obesity, and blood type), image data (such as brain MRI and lung X-rays), and genetic variants to assess the severity of COVID-19. We will employ and develop machine learning approaches, such as logistic regression, random forest, to predict the probability of COVID-19 infections, given the interpretability of potential clinical and practice characteristics.
2. As the non-interventions of controlling the COVID-19 have been issued for each infected country, we would statistically evaluate the outcomes of those interventions, such as blocking communities, closing the schools. The causal inference analysis will be employed to identify the treatment effects of the interventions. For example, whether the public opinion about the vaccine shot could influence the number of positive individuals of COVID-19.
By undertaking the above analyses, we aim to build awareness of the critical risk factors that influence the transmission of COVID-19, provide a reliable source and evidence for practical people, and create new and innovative solutions for the efforts made along the way.